Bishop Hill Try, try again
Sep 22, 2011 Climate: other Media Alexis Madrigal at the Atlantic reckons the problem with the whole climate change thing is just some nuances in the message. Polar bears are out and people are in:
The nugget of the argument here is the framing fighting climate change as a way to help nature is flawed. Even in a really clear example of climate-induced ecological change -- the loss of many species in the forests Thoreau explored near Walden Pond -- other species are doing just fine. We're changing the ecosystem, but life isn't leaving the forest. We're applying a very certain kind of filter on the forest (specifically, which plants can change their flowering time quickly) but life there survives because ecosystems, even those stressed by rising temperatures, are resilient.
Human-built environments, on the other hand, are very efficient and very brittle. They function best in a very narrow set of temperature and precipitation conditions. Witness what happens when it snows in Portland or it gets very hot in a cold place or it rains somewhere where it's always dry. A people as rich as Americans can deal with any climate, but only if we put the right infrastructure in place. People in Buffalo have snow plows. People in Phoenix have air conditioners.
This idea that the answer to the climate change problem is more media or different media is one that gets wheeled out every few weeks it seems, without anyone seeming to get it - it's not the way you are telling the story guys, it's the fact that the message is so obviously politicised that nobody trusts you.
Reader Comments (89)
"Personal opinion is out until we get this sorted."
Personal opinion ia always in.
Andrew
Spence-UK
Total hand-wave. References please.
Spence_UK
Also: what difference does it make? Imagine that your rather outre hypothesis is correct. We still have a minor change in forcing (DSW) and a major climate system response. Which absolutely requires a high climate sensitivity involving net positive feedbacks.
Address this directly in your response, please.
BA
You are one of the laziest posters here. Not to mention deliberately provocative and occasionally directly offensive. I am heartily sick of you.
Here's a suggestion: why not do as I invite above and find a referenced definition for 'scientific consensus'. Then, for perhaps the first time, you can make a substantive contribution.
Prove to us all that you aren't just a silly kid with a keyboard and too much time on your hands.
Go on.
Please take note of my earlier posting about conduct on conduct on comments threads.
"You can't tell the players without a scorecard" and you can't make sense of climate science without understanding the politics. Climate science IS politics. It can only be understood when viewed through a political template.
In the US right now, the left has degenerated to this really weird position where nothing matters but soundbites. We are in the age of Humpty Dumpty. A green job is anything they say it is (a bus driver for example -- he drove a bus for 30 years, but if govt helped fund the building of a hybrid bus which he now drives, his job is now green and counts as one created by the govt funding.) Spending cuts in laws passed 2 years ago are now considered part of a new proposal for scoring purposes. Medicare savings numbers are shamelessly double-counted if it makes the story sound better. Jobs were 'created' if the same model used to propose the stimulus is then used to "prove" they must have been created. It's a looking glass world.
It's all BS, all the time. Reality, which always took a back seat to perception in politics, has now been relegated to the garage. Spin and managed perception dominate. If the media will accept it and foist it on the gullible, it becomes the new truth. Slander, lies, and straw men are the coin of the realm.
If this is the standard for political communication, no one should be at all surprised that the political communication which happens to involve climate issues should be handled the same way as political communication involving health care, taxes, or social security. If the old putrid stew of sound bites being used to sell the nasty dog food is now polling badly, the obvious solution is to stir up a new stew of sound bites to sell the nasty dog food. Surely the public can be convinced to buy it, if only the right strategy of sound bites and BS can be concocted.
If polar bears and disappearing islands aren't working any more, substitute carbon 'pollution' and extreme weather events. If drought and floods and tornadoes and hurricanes don't work, they will go on to plan C, then D, then E. The political end game will never, ever change. Just the sound bites and BS thrown up to try to win public support.
BBD "That's why those arguing for a low CS (and I was one, for years, remember) are on a hiding to nothing."
I can't even begin to understand your logic here so I must be missing something. You seemed to be fixated on coming out of an ice age (and presumably going back in again) as though this gives you all you need to understand about climate sensitivity.
I just don't get it. Some kind of 'thermostat' is clearly at work (thus my earlier link to Willis' post) but as yet we don't know exactly how.
BBD-
"The problem is that this doesn't fit with the rapid jumps from ice ages to warm interglacials like this one. These are triggered by slight eccentricities in the Earth's orbit. The little push."
Indeed. We have ice core data that relies on a temperature proxy with still-unknown accuracy at a few point locations on the globe, 'robustly' teleconnected to a global climate state, that is then correlated exclusively with estimated but unmeasured forcings, to conclude that climate sensitivity is high. This is yet another exercise in Paleo-Rorschachism.
Meanwhile, measured temperature records, when compared with measured forcings, collected globally, repeatedly demonstrate a low climate sensitivity.
But not to worry, Gavin Schmidt at NASA GISS supports you concerning the consensus-approved 'imperfection-free' paleoclimate record, the data for which Steve McIntyre still has received nothing from Lonnie Thompson-
“Climate sensitivity is not constrained by the last two decades of imperfect satellite data, but rather the paleoclimate record.” Gavin Schmidt, 8/5/2011
But wait. There may be a problem with using ice cores or Bristlecone pines at a few locations on the globe-
8/21/2011
James Renwick, principal climate scientist at the National Institute of Water and Atmospheric Research, comments on snow storm in NZ-
“When you look at the globe, New Zealand is a tiny little patch on the Earth and to try and relate what is going on here, in terms of temperature, to global change is really meaningless.”
Surely the words from this learned climate expert apply to all sparsely located proxies?
BBD -
May I suggest opening a discussion thread on the topic of climate sensitivity from the perspective of interglacials? It's come up in several threads recently. I'd start one myself, but I think the thesis statement should be yours. Sounds like Spence_UK would contribute, and I have some thoughts as well.
http://washingtonexaminer.com/opinion/2011/09/judge-questions-honesty-interior-department-scientists
Start with the policy preference. Science becomes whatever is needed to support the policy sought.
BBD, I'm happy to furnish references; since you hadn't either I didn't realise we were going to that depth.
Assuming you already have some knowledge of HK dynamics, this link is probably most appropriate (ref 1).
If you need some background on HK dynamics, this link provides some background (ref 2).
Note I've linked to presentations and lectures here, the papers upon which they are based are cited within.
ref 1. Markonis, Y., D. Koutsoyiannis, and N. Mamassis, "Orbital climate theory and Hurst-Kolmogorov dynamics", 11th International Meeting on Statistical Climatology, Edinburgh, International Meetings on Statistical Climatology, University of Edinburgh, 2010.
ref 2. Koutsoyiannis, D., "A random walk on water" (Henry Darcy Medal Lecture), European Geosciences Union General Assembly 2009, Geophysical Research Abstracts, Vol. 11, Vienna, 14033, European Geosciences Union, 2009.
IIRR BBD's response will be along the lines of "Is HK mutually exclusive with AGW?"
Spence_UK
Thanks for the refs. I thought Koutsoyiannis might come into it somewhere. Before we vanish down that particular rabbit hole, can I repeat my second question:
Also: what difference does it make? Imagine that your rather outre hypothesis is correct. We still have a minor change in forcing (DSW) and a major climate system response. Which absolutely requires a high climate sensitivity involving net positive feedbacks.
Address this directly in your response, please.
ALL
Please continue on Discussion thread... ;-)
BBD: two points.
One. HK dynamics is NOT a response to forcing as you characterise it here. You have not understood the references I have already provided you if that is your response. It is a consequence of the internal dynamics of a complex, coupled, non-linear system. The power spectral density we see in the proxy records at the 100kyr cycle length is consistent with no external forcing at all.
Two. I have no idea how one scientifically defines what an "outre hypothesis" is. As noted above, HK dynamics is the ONLY model consistent with the power spectral density of the orbital forcing, and is also consistent with the power spectral density of everything from satellite temperature data to geological timescale proxies. No other single model achieves this without kludges. To me, a model that fits the evidence well is a good model. To you, I guess it is an "outre hypothesis". OK.
Spence_UK
No, and that is the problem. I do not accept that the unforced internal dynamics of the climate system can terminate a glacial.
Koutsoyiannis' work is always interesting, and looks where others do not, but I'm not persuaded on this point.
Spence_UK
Nor does this address the issue of climate sensitivity. For internal variation to be capable of causing huge climate shifts, net feedbacks must be strongly positive.
If you disagree, then explain the mechanism by which internal variability causes major climate shifts.
If you believe that CS is low, then this must be incorporated into your explanation.
Firstly, having looked at the ordered statistics of the past 650kyrs, I see no evidence of two distinct states, but a continuum of variability. I think the idea of "terminating a glacial" implies a (false) dichotomous condition of the earth's climate, imposed via a (somewhat) arbitrary threshold though human analysis. I don't see a point to it.
The magnitude of temperature swings we have seen in the past 650kyrs is completely consistent with the power spectral density we see from weather scale (month to month, year to year), all proxies I am familiar with and a fixed Hurst exponent in the region of 0.9 to 0.95. Nothing else is needed.
If I were to come along and say, "I can't see El Nino switch to La Nina without some kind of forcing", would you find that an odd comment? To me, there is no difference between this and your claim regarding the glacial "terminations". Both can be explained by natural variability within the climate system. Both fit well within the HK dynamics model and require no explicit external forcing. (Indeed, a forcing on this scale up to the magnitude of the HK dynamics would be unmeasurable).
Sorry, this one crossed over my last posting and I missed it.
Lots of terms here which I do not think have meaning. What is a "huge climate shift"? HK dynamics present a continuum of change consistent with the observations. Different scales exhibit different magnitude changes linked via a simple relationship. "Huge shifts" is not clearly defined in this context.
Complex non-linear dynamics bounded by simple constraints and an assumption of entropy maximisation; this is the mechanism outlined in the paper "a random walk on water" outlined above.
You are changing your original argument here. The argument I am addressing is your original one, in which you explicitly claimed above a low CS could be ruled out as there was no explanation for the glaciations other than a high CS. I am pointing out that there is an alternative explanation quite consistent with a low CS. This does not "prove" low CS, but it means you cannot rule it out as you did above. That is what I am questioning.
However, I see no point in invoking high CS where no reason to invoke it exists. Neither models, observations nor proxies require high CS to be invoked to explain them.
"A people as rich as Americans can deal with any climate, but only if we put the right infrastructure in place. People in Buffalo have snow plows. People in Phoenix have air conditioners."
Unfortunately, the AGW types - Hansen, Mann, Ban Ki Moon, All Gory... - say that we must get rid of snow plows because they are made using and in turn use "fossil fuel," and we must get rid of air conditioners because they "use too much power" and/or "fossil fuel power plant" power. Alexis Madrigal is probably among them.
These are also the people - e.g. Hansen - who demonstrated (and threatened actual violence) when E.On started to build a bigger fossil-fuel-plant - which was needed to supply power during the fifty-seventy percent times the huge "government paid/subsidised" (by everyone's taxes and electricity rates) wind farm they were also building next door would not supply power (no wind, too much wind, too cold, too hot).
Spence_UK
So glacials (Ice Ages) and interglacials are a false dichotomy :-)
Disagree entirely. You need to back this up.
GAT increase of 5C in a century or less.
I've changed nothing. You have not explained how internal climate variability can have large climatic effects if CS is low. How does it work? Entropy maximisation? What do you mean by this? How does it work in terms of climate behaviour? And how does it square with a low CS where energy is efficiently shed to space?
Yes. The proxy records (say the 650kyr Vostok ice core d18O temperature reconstruction) show a continuum of variability with no distinct steps. I already explained how to check this. Simply sort the proxy data and plot temperature against sort index. If there are distinct states, they will appear as ledges on this plot. What you find is a pretty much straight line. Nature does not distinguish between two states, there is a continuum of temperature and a continuum of ice cover. The self-similar nature of the time series tends to cause these to cluster, and human analysts create a false dichotomy by applying an arbitrary threshold (which does not exist in nature) to distinguish two "states". But these states are merely an arbitrary threshold on a continuous variable, not true states. Hence the false dichotomy.
I'm biting my tongue on this a little bit - given you are backing up none of your claims yet demanding I back mine up - but I'll carry on for now. Many of the Koutsoyiannis lectures back this up, but it can be seen most clearly on pp 28-29 of one of his most recent presentations (see here, ref 1).
What is the basis of this number? Why do you think it is anomalous? We've seen swings of about 1 deg C within a year from internal variability in the satellite era, why should we exclude 5 deg C in a century from unforced variability? What mechanism are you invoking that prevents it?
In fact, yes I have. I have linked to papers clearly explaining it. The physics behind it is not difficult, but needs some effort on the part of the reader to follow. You need to clear certain conceptual hurdles before it makes sense. I can give some simple analogies to try to help understand, but bear in mind they are analogies and not the full story.
The hydrological cycle probably has the biggest impact on temperatures. Just look at cloud cover; have you ever noticed that clouds have a pretty huge effect on the temperature? A clear night or cloudy day will be colder than normal; a clear day or cloudy night warmer.
But cloudiness tends to cluster. If it is cloudy one hour, it is more likely to be cloudy the next. If there are clear blue skies, it is likely that the day will be mainly clear. Not guaranteed, just more likely. This clustering is due to the fractal nature of cloudiness.
But when we analyse data, we see this clustering doesn't just occur at the daily level. It extends out to months, years, decades, centuries... and the evidence we have suggests out to very long scales. And not only temporally, but spatially as well. The consequence of this clustering is that the local sample mean is a poor estimator for the population mean.
This is nature's own internal variability. It isn't forced, it's a part of nature. When you study the time series you see this behaviour in all manner of metrics - from cloudiness, to temperature, to ocean currents. We see it in ENSO (as alluded to above), which in 1998 saw a swing of around 1 deg C from internal variability over the space of twelve months.
Understanding how this variability scales is critical. What we see in well dated proxies (see diags from 28-29 in ref 1) is that the variability consistently expands from increasing scale. Professor Koutsoyiannis demonstrates in ref 2 why this is expected from the principle of maximum entropy with some simple constraints. Why is this more compelling? Because it explains climate patterns far better than the deterministic example. The 100kyr anomaly I discussed above is just one example of where the deterministic forcing model fails to match the data. Yet the 100kyr variability is exactly where HK dynamics says it should be.
ref 1. Koutsoyiannis, D., Hydrology and Change (Plenary lecture), IUGG 2011, Melbourne, International Union of Geodesy and Geophysics, 2011.
ref 2. Koutsoyiannis, D., Hurst-Kolmogorov dynamics as a result of extremal entropy production, Physica A: Statistical Mechanics and its Applications, 390 (8), 1424–1432, 2011.
Spence_UK
This is not true. You are saying stuff. Back it up.
You make an extraordinary claim. Extraordinary evidence is required to support it. I see none.
Here are lots of pretty pictures of that 'continuum of variability with no distinct steps'.
Pick your favourite.
BBD - thanks for that. This was the second one I clicked on and it is definitely my favourite:
vostok-cores-show-zero-climate-sensitivity"
( http://stevengoddard.wordpress.com/2011/06/19/vostok-cores-show-zero-climate-sensitivity-2/ if I not get the html link right)
BBD, I'm disappointed in your responses to be honest.
What you see as structure in the time domain is anything but. They are normal variability for a distribution with fractional integration. You're not familiar with this stuff so you try to explain this structure but what you see are false dichotomies. I explained why they are false dichotomies and how you can see this for yourself but it seems you can't see the wood for the trees.
You argued that there was some kind of transition between two states that could not happen without forcing. That nature somehow sits in one of these states - "glacial" and "non-glacial" and that natural variability could not switch between them without some kind of external forcing.
So I made a suggestion - a suggestion which you have clearly not done - to illustrate why this is a false dichotomy, to help you understand. Take the vostok data and sort it. If there are distinct states, they will show up as ledges on any such plot.
A straight plot of the deuterium results (which have more samples and therefore detail than the d18O, but both give the same conclusion) shows a continuum of points, with a slight bulge near 0. The bulge is simply because the samples are non-uniform, with many samples during the recent past (at multi-decadal to centennial resolution) whereas the bulk of the record is nearer millenial resolution. Truncate those points and you get a very clear, continuous curve. The pic is linked here.
(The chart with distortions from temporal resolution is here)
The sorted data is, in effect, a cumulative density function of the distribution of temperatures. The shallower the line is, the more probable those temperatures are; the steeper, the less probable. The graph shows a CDF that is clearly continuous and looks unimodal, albeit slightly asymmetric, with no evidence of preferred states. Now, you suggested that there was a preferred state ("glacial") and that this state could not be "terminated" (let's call it non-glacial) by natural variability. If that is true, you should be able to draw a horizontal line on my CDF pointing out where this transition from glacial to non-glacial occurs, and you then need to evidence your claim that there is something that prevents natural variability crossing from one side to the other side of that horizontal line.
And I say there is no line that can be drawn on that graph that cannot be crossed by natural variability. The apparent structure you see in the time series is simply random clustering that is fully explained by Dr Koutsoyiannis' work.
As for evidencing the clouds, those were an analogy to help you understand the principles, not the evidence. I've already linked you to the evidence clearly laid out in the peer-reviewed literature and in presentations and lectures. If you're not willing to put the time and effort in to understand it (and it isn't exactly something you can pick up overnight anyway), then I can't help you. The evidence is clearly laid out. Simple analogies have been drawn to help understand. I'm not sure what else I can do. Understanding is not an algorithmic process.
Also: on checking the deuterium data, the biggest calibrated century-scale temperature step is just over 2 deg C. So I have not got a clue where your 5 deg C GAT in 100 years step comes from (bearing in mind vostok is, at best, a SH reconstruction).
Perhaps you are thinking of the 8.2kyr event; but as the vostok ice core shows, this was hardly 5 deg C globally. So your "major climate shift" doesn't even seem to be something we have credible evidence for. And that still ignores the fact that you haven't explained why it can't happen from natural variability alone; you've just asserted that it can't. Assertion is not evidence.
Sensitive to what? Is there any reason climate can't be differently sensitive to all its different impulses?
===============
the epistemology of scientific practice is an interesting one.
Science is an amalgam of models upon which mathematics is then unleashed, to forecast behaviour which can then be observed.
One can try to learn these models and the maths, at which time one is just a small rat on a treadmill,
Or one can "practice science", at which moment you are disputing either the models or the maths or the observations.
The former is a futile student , the latter is a sceptic. None is carrying "consensus".
the WSJ has an interesting article on tommie friedman, one of the paragons of the Loose Nut Left, who proclaims, rather amusingly, that America is at war with maths and sciences. He does not know what either is, of course, but the statement is an interesting one.
He must refer to the publicised emails trove of the UEA CRU ?
Spence_UK - the sort test is a nice idea, I don't have the data and don't have time to search for it. Maybe I have seen it in the past and forgotten but I'd like to check your assertion that you arrive at a straight line?
I'm interpreting your test as a simple ranked plot of temps. So each data point on the x axis takes its rank position in accordance with the temperature associated with it. This makes it independent of the "time" it occurred (unlike the plots BBD linked to) and it then tells us something about the distribution of temps. As you say this will help investigate whether or not states exist or if there is simply a continuous transition through a range of temps.
What I'm less sure about is if you will get a straight line? Wouldn't that only occur if every temp occurred an equal number of times? Looking at the temps vs. time plots I find this unlikely? Not far off admittedly, but if you have an example plot it would be good to see it (or a reference).
Thanks for any further info.
not banned yet,
The plot and description is shown in my Sep 24, 2011 at 1:43 AM comment. It is linked here:
http://i56.tinypic.com/24b9apz.gif
Yes, what you get is a cumulative distribution function which has no kinks, ledges or sudden changes - straight line was probably not the best description (as you note a straight line would be the CDF of a uniform distribution). A smooth, continuous curve with no evidence of preferred states is a better description of what I was getting at.
Because the curve lacks kinks, ledges or sudden changes, it seems reasonable to assume the distribution is unimodal rather than bimodal, i.e. that there are no particular "states" which nature preferentially selects, just a continuous variation of temperature.
The distribution is slightly asymmetric, and has a non-zero third moment.
This is the basis on which I note that the dichotomy of "glacial" and "interglacial" may have meaning to humans, in which we tend to categorise these things for convenience, but nature has no concept of these; just a continuously varying temperature and ice extent. There are no artificial barriers to any particular value.
Thanks - follow up appreciated.
Spence_UK
You are relying on a single researcher. I disagree that you have provided any good evidence in support of your position that unforced natural variation can terminate a glacial. I am not alone in this. You have allowed yourself to be persuaded on the flimsiest basis and in the face of considerable evidence to the contrary. This is unwise.
Next, and I have to say this in bold, the inappropriate application of statistical techniques to climate data has caused problems before. I urge caution.
The new understanding of the extraordinarily rapid climate shifts at the end of glacials is based on the Greenland ice core data. For example, see here:
And here..
As for this:
Well, where did the energy come from then? You are asserting that natural variability is the cause, but you have provided no evidence for the mechanisms despite my twice asking for some. I don't think you are in a position to start getting snippy about 'assertion is not evidence'.
Finally, the original point of this conversation was that any abrupt climate change (and glacial terminations are abrupt, whatever you claim) requires a relatively high climate sensitivity. You have not addressed this point.
Wow, BBD. I'm embarrassed that you would make such a poor argument. Relying on one researcher? Really? This is how we debate science now, is it? I'll show now why that comment is flawed on two levels.
Firstly, let's just assume that there really is only one researcher. So what? The ideas are published, and have not yet been shown to be wrong. On that basis alone, you cannot reject low sensitivity on the basis of the last few hundred thousand years. Because there is a valid, unfalsified theory that is compatible with a low climate sensitivity and the variations we have seen in these proxies.
So even if it was just one researcher, your point would be worthless. But here's the second point: it isn't just one researcher. Aside from the references I have already provided have more than one author, aside from the fact that the Itia group at NTUA has 20+ researchers, many others also support this viewpoint and have published supporting papers: including the late Vit Klemes, Dr Tim Cohn from the USGS, Dr Alberto Montanari, Dr Harry Lins, and so on. Still not enough for you? Even Dr Carl Wunsch published a peer-review article making the exact same point (albeit in a different way) warning people against reading too much into these proxy records. His paper is here, ref 1.
Your claim about just one researcher is not only anti-science at its very core, it is also flat wrong.
Moving on to Greenland. Really? You say that Greenland ice cores are good for examining entry and exit of glacials? Plural?
Do you even realise that Greenland was largely ice free during the Eemian? So even if we accept the human imposition of states on the ice ages, we only have one exit (the Eemian, which was very gradual) and the entry to the Holocene.
Furthermore, Greenland is well known for larger fluctuations that elsewhere. The Greenland ice cores show a very prominent MWP, RWP and GWP. Yet the same scientists who insist we use Greenland ice cores to detect highly sensitive swings during the ice ages disavow the exact same data when it comes to recent warm periods. This is what I mean by "kludges". This data means X when we want it to mean X, and the exact same data means NOT X when we want it to mean that. Science doesn't work like that; never has done, never will do.
As for the energy comment, where does the energy come for the positive feedbacks? The same place as for internal variability: the sun. The sun produces a continuous stream of energy, and changes in radiative properties (clouds, ice cover, etc) modify the temperature of the earth. This is the same for both cases, so is a non-argument. The climate sensitivity argument assumes a linear response from a complex coupled system. The stochastic argument assumes non-linear dynamics change the balance is a more complex way.
Or to put it another way: where do you think the energy comes from when ENSO changes the temperature by 1 deg C over 12 months, as it has in the satellite era? And why do you think 10 deg C over 10,000 years (which is a more realistic measure of variability over the scale of hundreds of thousands of years) is so difficult? That is just 0.1 deg C per century. Nothing compared to what ENSO achieves.
It is trivial on the face of it. You just change the cloud cover and the ratio of sunlight changes, and the earth's energy distribution changes. This isn't rocket science. It is trivial and obvious. But cloud cover doesn't change in a linear way, and the point the Itia group has been making is that scale averaging does not bring the sample average closer to the population average.
ref 1: Wunsch, C., "Quantitative estimate of the Milankovitch-forced contribution to observed Quaternary climate change", Quaternary Science Reviews 23 (2004) 1001–1012
And while I remember...
Yes. And this is EXACTLY the point being made here. Such as creating false dichotomies where none exist, and hand-waving about 5 deg C GAT from a regional proxy that was already disavowed from the MWP.
There is only one kludge-free model that captures all of this variability. And it isn't the one you're selling here.
Spence_UK
Oh no! Something else the whole field has got wrong :-(
But wait... what's this?
Once again, take it to the discussion forum.
So, you don't understand what the word "largely" means either?
Tell me then. How much data have we got from ice cores prior to the Eemian about the climate of Greenland?